All these new servers/users/followers are filling my poor little single-user instance up with data fast. There's so many of you!
Already knocked myself offline trying to move everything to S3, and failing at it.
I know what I did wrong, so I'll be trying again later today after sanity checking my process.
I can only imagine what it must be like to run a big instance. o7 to the community admins out there. Thank you for your service.
@Alex I recently moved my file store to S3 - found a pretty good explainer. Need a link?
@kevin Yes please! I've been following the cybrespace instructions, but I ran into a URL caveat.
Though it might because I am using DigitalOcean Spaces (S3-compliant).
The guide I am using: https://github.com/cybrespace/cybrespace-meta/blob/master/s3.md
@Alex oh that's the exact guide I used!
What URL caveat have you run into?
@kevin After migrating all my data, I found that the files are expected to be stored in
'bucket.region.digitaloceanspaces.com/bucket'
instead of just
'bucket.region.digitaloceanspaces.com'
No idea why, but moving all 150k files one directory deeper kept failing, so I restored from the snapshot I took before I made changes, and plan to start over from scratch with that caveat in mind.
@Alex Mine are not under bucket.region.aws.com/bucket, they're in the root folder.
I used the single s3cmd in the cybrespace instructions and didn't try to create multiple threads processing the copy
@kevin Weird... I wonder why I had different results?
@Alex No idea. Is the Mastodon app actually seeking them in the /bucket/ structure and 404ing?
@kevin Yeah, I migrated all my files to the root the way the docs said to, and after changing my config and restarting the service, all the images were broken because they contained the bucket name as the parent folder.
Doing a right-click, Open Image, and removing the bucket name from the directory tree took me to the appropriate files.
My next worry is that new files are going to end up in the root and further break things when I try again.